Concept learning is learning about the regularities or patterns that separate stimuli into different groups, or concepts. The overall aim of this project is to differentiate between forms of concept learning and their constituent subprocesses through the identification of participating neural systems using functional Magnetic Resonance Imaging (fMRI). The first aim is to dissociate the neural substrates of rule learning (classifying on the basis of a verbalizable rule, such as "always choose the blue stimulus") from exemplar learning (classifying on the basis of similarity to previously learned stimuli), and to compare the brain activity of learners with persons who fail to learn. The second aim is to dissociate the neural systems underlying rule formation from rule application. Rule learning is dependent on executive functions subserved by the frontal lobes, and thus these studies have the potential to elucidate the pattern of deficits seen in patients with frontal lobe damage due to stroke or traumatic brain injury. The third aim is to related recruitment of striatal brain structures in concept learning to the degree that a probabilistic relationship exists between a stimulus and its concept membership. Because the striatum is damaged in disorders such as Parkinson's and Huntington's diseases, this study may provide insight into the types of learning problems seen in patients with these diseases.